zipper_algebra: further specialize grafting

[marcin-rzeznicki]

This leads to a few percent speed-up in the benchmarks. For instance, for symmetric difference simulation, before this we had:

ops               fastest       │ slowest       │ median        │ mean          │ samples │ iters
╰─ xor                          │               │               │               │         │
├─ base_2b     84.04 ms      │ 85.36 ms      │ 84.28 ms      │ 84.47 ms      │ 5       │ 50
├─ base_4b     888.8 ms      │ 896.3 ms      │ 893.7 ms      │ 893 ms        │ 5       │ 50
├─ zipper2_2b  494.2 ms      │ 501 ms        │ 497.1 ms      │ 497.9 ms      │ 5       │ 50
├─ zipper2_4b  3.88 s        │ 3.897 s       │ 3.886 s       │ 3.887 s       │ 5       │ 50
├─ zipper4_2b  331.4 ms      │ 336.5 ms      │ 333 ms        │ 333.7 ms      │ 5       │ 50
╰─ zipper4_4b  2.4 s         │ 2.411 s       │ 2.403 s       │ 2.404 s       │ 5       │ 50

and now we have:

Timer precision: 33 ns
ops               fastest       │ slowest       │ median        │ mean          │ samples │ iters
╰─ xor                          │               │               │               │         │
├─ base_2b     71.11 ms      │ 72.15 ms      │ 71.48 ms      │ 71.55 ms      │ 5       │ 50
├─ base_4b     838.6 ms      │ 858.7 ms      │ 840.8 ms      │ 844.1 ms      │ 5       │ 50
├─ zipper2_2b  308.1 ms      │ 312.8 ms      │ 310.3 ms      │ 310.2 ms      │ 5       │ 50
├─ zipper2_4b  2.625 s       │ 2.65 s        │ 2.635 s       │ 2.636 s       │ 5       │ 50
├─ zipper4_2b  158.1 ms      │ 159.3 ms      │ 158.7 ms      │ 158.7 ms      │ 5       │ 50
╰─ zipper4_4b  1.603 s       │ 1.617 s       │ 1.615 s       │ 1.612 s       │ 5       │ 50

so there is a quite considerable speed-up (especially for longer paths)

[adamv-symbolica]

Let's fold do-graft optimization into the masked version @marcin-rzeznicki